Podcast Summary

Work for Humans

Episode: AI as Dramaturg: What It Means to Create Art with a Machine
Host: Dart Lindsley
Guests: Matthew Gazda (playwright), Isabel McCrum (Microsoft language scientist, theater producer)
Date: October 7, 2025

Overview

This episode explores the intersection of artificial intelligence (AI) and human creativity, specifically through the lens of theater. Playwright Matthew Gazda and producer/AI scientist Isabel McCrum join host Dart Lindsley to discuss their collaboration on Gazda’s play Doomers, which dramatizes the tumultuous weekend when Sam Altman was nearly ousted from OpenAI. The conversation investigates AI’s role as an artistic collaborator ("AI as dramaturg"), the uneasy reactions of audiences when AI is credited creatively, and the philosophical boundaries of authorship, depth, and meaning in art created with or by machines.

Key Discussion Points and Insights

1. Origin of the AI-Art Collaboration

• Matthew Gazda’s Playwriting Process:
  • Inspired by real-life tech drama involving Sam Altman, Gazda sought to depict Silicon Valley culture authentically, consulting with AI professionals, and increasingly, with AI itself.
  • Quote: “There’s something about that night... speculating about OpenAI and Sam Altman's firing... very dramatic, very tragic, a kind of sense of the hero having a tragic flaw that leads to his potential destruction.” —Matthew Gazda [04:04]
• Embedding himself in the tech world, Gazda used AI tools not just for research, but to mimic the real communication styles of Silicon Valley insiders.

2. AI as Dramaturg: What Does That Mean?

• Traditional Role of a Dramaturg:
  • Described as an "in-house philosopher" or "intellectual assistant," usually providing critique, fact-checking, and boundary-setting, but rarely authoring content directly.
  • Quote: “It's basically an in house critic... someone who thinks about what the play means... a kind of intellectual assistant or referee, boundary setter.” —Matthew Gazda [12:27]
• AI’s Function in the Process:
  • Used AI language models (Claude, ChatGPT) to check plausibility and authenticity of dialogue for various professional and cultural perspectives.
  • In Doomers, an AI-generated allegory by Claude appears verbatim as a character's monologue—demonstrating a literal, multifaceted contribution.

3. Audience Reactions and Human Anxiety

• Fear of Alien Collaboration:
  • McCrum recounts audience uncertainty and fear upon seeing AI (Claude and ChatGPT) credited in the playbill—symbolizing the "alien-ness" and mystery of AI’s creative role.
  • Quote: “The fact that this alien thing was involved in the creation of the play only inspires fear in that moment.” —Isabel McCrum [11:30]

4. Experiment: Training an LLM on Gazda’s Work

• McCrum’s Experiment:
  • Fine-tuned Google's Gemini on Gazda’s plays ("MattGPT") to generate AI-authored outputs, intending to probe both authenticity and audience perception.
  • Initial idea: Stage two versions of the play (human- and AI-written) and observe audience emotional responses and ability to distinguish them.
  • Quote: “I wanted the model to follow his style of play exactly... If it does generate something that is good, what does that mean?” —Isabel McCrum [17:54]

5. Parody, Reflection, and Limits of AI-Generated Art

• Results of the Experiment:
  • AI often produced parodies of Gazda’s style—useful for self-reflection but rarely subtle or truly original.
  • LLMs can highlight an author’s verbal tics or clichés, functioning as a self-parody mirror.
  • Quote: “It did come back as a parody. Very much so, for the most part.” —Matthew Gazda [21:03]
  • Quote (on reflection): “Parody does tell you a lot about your own weak points, or at least your strengths and weaknesses.” —Matthew Gazda [21:10]
• AI can help brainstorm plot points or offer disruptive linguistic inputs, but fails to capture deep human nuance and subtext, especially in dialogue.

6. Depth, Subtext, and the “Plastic” Problem

• Human vs. Synthetic Expression:
  • AI-generated text is “plastic”—even if indistinguishable on the surface, it lacks the perceptible depth or lived experience of human-authored art.
  • Lindsley: “There is no Straussian reading of an AI. There is no subtext behind the AI. It's like a lack of depth.” [32:30]

7. Borges’ Pierre Menard Thought Experiment

• Discussion of Borges' story, Pierre Menard, Author of the Quixote, where the source/author matters as much as (or more than) the text itself—even if two works are identical.
  • Quote: “There's a non zero chance that Isabelle's LLM does produce doomers, or at least a couple pages that are identical with my play...and the answer is if the actors knew that it was written by [AI], they almost certainly would interpret the text differently.” —Matthew Gazda [34:20]
  • “We consider these pieces of art within a tale or within a larger context... why does that matter to us so much? Because it does. It really, really does.” —Isabel McCrum [36:21]

8. Originality, Aura, and Audience Experience

• Conversation about why originals matter in art (even when perfect reproductions are possible), tying into questions of human touch, intent, provenance, and “aura.”
• Quotes:
  • “If you convinced me that it was an original Picasso, I would be probably just as happy as I would be if I didn't. But yet I'm still not completely convinced that there isn't something that happens with matter when it's touched by people.” —Matthew Gazda [38:40]
  • “Life as training data.” —Dart Lindsley, reframing human experience in algorithmic terms [41:21]

9. Context, Audience, and the Mystique of Liveness

• Live theater’s intimacy and the sense of community and uniqueness are part of its "original art" aura—undermined if AI co-creates or automates the process.
  • “Your plays in particular... have that feeling of original art because they are often in very small venues, in people's apartments originally. So there's a sense in which I am uniquely experiencing a piece of art.” —Dart Lindsley [43:05]

10. The Future: Blended Roles and Changing Definitions

• Discussion of how technological changes—like photography's impact on painting—shift artistic values and audience perceptions, hinting at the unpredictable future of AI in art.
• Quote: “Art might just be the best pattern matching in the right time and place... the context around my work changed.” —Matthew Gazda [49:00]

Notable Quotes & Memorable Moments

• On AI’s role in art:
  “Mind mesh in the play is a character who... prompts mind mesh in the course of the play and gets this response. And I used... the story that Claude gave me was so weird and kind of chilling. I was like, I have to actually use this.” —Matthew Gazda [14:34]
• On the reaction to AI’s presence:
  “The fact that this alien thing was involved in the creation of the play only inspires fear in that moment.” —Isabel McCrum [11:30]
• On the difference between AI and human authorship:
  “If the dramaturg starts actually stepping in and writing large sections of it independently... pretty soon, it's not Matt at all. It's Matt's prompt, I guess.” —Dart Lindsley [16:17]
• On subtext and authenticity:
  “There is no Straussian reading of an AI. There is no subtext behind the AI. It's like a lack of depth.” —Dart Lindsley [32:30]
• On the context of performance:
  “If Rodin makes his clay model and then you receive the cast in bronze... at what point does it stop being Rodin's thinker?” —Matthew Gazda [44:09]
• On art as synthesis:
  “Art might just be the synthesis of what we believe about art and what it really is at all times.” —Matthew Gazda [48:01]

Important Timestamps

• 00:03 — McCrum recounts overheard audience fears about AI authorship
• 04:04 — Gazda describes origins of Doomers and his entry into AI and tech culture
• 11:30 — McCrum on the “alien-ness” of AI collaboration in the arts
• 12:27 — Gazda defines dramaturgy and elucidates AI’s function as “dramaturg”
• 17:54 — McCrum explains fine-tuning an LLM on Gazda’s plays (“MattGPT”)
• 21:03 — Results of the AI experiment: parody, not subtlety
• 26:32 — Discussion of LLMs as creative disruptors, mirrors, and references
• 32:30 — On subtext and the absence of depth in AI-authored art
• 34:20 — Borges’ Pierre Menard parallel: the importance of origin in evaluation of art
• 38:40 — Why originals matter, even if visually indistinguishable from copies
• 41:21 — “Life as training data” and human existence as algorithm
• 44:09 — The role of context, audience intimacy, and the “aura” of liveness in original arts
• 48:01 — The evolving, social nature of art and authorship
• 55:51 — What do you hire your work/job to do for you? (Closing reflections)

Tone

The conversation is probing, intellectual, candid, occasionally humorous, and deeply reflective—balancing practical anecdotes with philosophical speculation. The guests’ openness, critical perspective, and lively banter with Dart Lindsley create a thoughtful, accessible exploration of art, technology, and the perennial quest for meaning and originality.

Conclusion

This episode is a must-listen for those interested in creativity, the philosophy of art, and the practical and existential dilemmas posed by AI in human domains. Gazda and McCrum’s play with the boundaries—sometimes literally—of dramaturgy, authorship, style, and audience perception, invites listeners to grapple with the rapidly evolving meanings of “art,” “human,” and “collaboration” in the age of intelligent machines.